In a communications network, large numbers of components such as computers, workstations, or file servers, are electrically connected by a communication network technology such as Ethernet, asynchronous transfer mode (ATM), fiber distributed data interface (FDDI), a technology known as TP-PMD (a copper-wire derivative of FDDI), and a networking technology known as 100VG-AnyLAN, which uses an access method called demand priority access method (DPAM). An Ethernet or other communication network typically includes a hub which is connected to the arrangement of components by communication cables, and which allows the computers, workstations, or file servers to exchange data signals. Data signals sent from a transmitting component to a receiving component are transmitted to the hub and repeated at the hub for transmission to the receiving component. The hub enables multiple computers, workstations, or file servers to share resources in a variety of applications. These applications include client-server database systems, in which a back-end database "engine" handles queries from multiple client front-ends running on desktop personal computers. The volume of data carried over the communication network escalates considerably as new users, new applications software, and more powerful computers or workstations are added to the network. As the volume of data carried over the network increases toward the maximum capacity, the data transfer rate through the hub and communication cables decreases, causing delays in computer applications and severely reducing the effectiveness of the network. Further, as the number of users associated with a network increases, more access ports are needed. To alleviate this problem, it is highly desirable to increase the capacity and/or the speed of the network.
A typical network hub includes one or more devices for routing data transfers between a number of ports (e.g., 12) in a workgroup. Each port may be assigned to one or more individual users or one or more individual computers, workstations, or servers. To increase the number of ports available to a workgroup, multiple hubs may be connected. Hub connections are typically achieved by uplink cables, such as unshielded twisted pair (UTP) cables, shielded twisted pair (STP) cables, or fiber optic cabling. In large, complex networks, a significant number of cables may be required. Cables present significant design limitations. For example, the total length of cable between hub units in a high-speed (e.g., 100 megabits per second) network must be less than 205 meters, and the total length of cable from a hub unit to a computer or other component must be less than 100 meters. Further, cables cause signal delay which can contribute to delays in network applications; thus, longer cables cause increased delay. In addition, signal reflection occurs at cable termination or connection points; thus, an increased number of cables causes increased delay. The reflected signals at the cable termination points contribute to signal degradation and inhibit network performance.